Vertex AI Vulnerability Exposes Google Cloud Data and Private Artifacts

Cybersecurity researchers have disclosed a security "blind spot" in Google Cloud's Vertex AI platform that could allow artificial intelligence AI agents to be weaponized by an attacker to gain unauthorized access to sensitive data and compromise an organization's cloud environment. According to...

PT-2026-32078

Product: Nuxt OG Image Version: injection via html parameter GET / og/d/og.png?html= When verbose errors are enabled, the response content is leaked in base64-encoded error messages. Vector 3: SVG injection via html parameter GET / og/d/og.png?html= Mitigation Fixed in v6.2.5. The image source...

Malware and Ransomware Detection in M365

Availability Requirement Threat Detection is available to Veeam Data Cloud for Microsoft 365 customers with Premium or Advanced plans. Customers must opt in to AI settings to enable this feature. Contact your Veeam account team or see your plan details to confirm availability. Supported Workloads...

CVE-2026-34204: Improper Authentication

MinIO is a high-performance object storage system. Prior to version RELEASE.2026-03-26T21-24-40Z, a flaw in extractMetadataFromMime allows any authenticated user with s3:PutObject permission to inject internal server-side encryption metadata into objects by sending crafted X-Minio-Replication-...

CVE-2026-34740: Server-Side Request Forgery (SSRF)

WWBN AVideo is an open source video platform. In versions 26.0 and prior, the EPG Electronic Program Guide link feature in AVideo allows authenticated users with upload permissions to store arbitrary URLs that the server fetches on every EPG page visit. The URL is validated only with PHP's...

PT-2026-29311

Name of the Vulnerable Software and Affected Versions Discourse versions 2026.1.0 through 2026.1.2 Discourse versions 2026.2.0 through 2026.2.1 Discourse versions 2026.3.0 through 2026.2.9 Description Discourse, an open-source discussion platform, had a flaw where non-staff users could view read...

CVE-2026-34163: Server-Side Request Forgery (SSRF)

FastGPT is an AI Agent building platform. Prior to version 4.14.9.5, FastGPT's MCP Model Context Protocol tools endpoints /api/core/app/mcpTools/getTools and /api/core/app/mcpTools/runTool accept a user-supplied URL parameter and make server-side HTTP requests to it without validating whether the...

PT-2026-29263

FastGPT is an AI Agent building platform. Prior to version 4.14.9.5, FastGPT's MCP Model Context Protocol tools endpoints /api/core/app/mcpTools/getTools and /api/core/app/mcpTools/runTool accept a user-supplied URL parameter and make server-side HTTP requests to it without validating whether the...

PT-2026-32059

Summary In openssl encrypt/modules/json validator.py at lines 234-238, when the jsonschema library is not installed, all schema validation is silently skipped with only a print warning. Affected Code python if not JSONSCHEMA AVAILABLE: printf"Warning: Cannot validate against schema 'schema name' ...

PT-2026-44334

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description An issue exists in the msm ioctl gem info get metadata function where it unconditionally returns 0, ignoring error codes stored in ret. This occurs when the user buffer is too small or...

Discourse Information Leakage Vulnerability

Discourse is Discourse open source set of open source community discussion platform. The platform includes features such as community , e-mail and chat rooms . Discourse suffers from an information disclosure vulnerability that stems from non-employee users having access to read receipt informati...

PT-2026-33866

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.3.28 Description An authorization bypass exists in the 'chat.send' gateway method. The issue occurs because ACP-only provenance fields are gated by self-declared client metadata from the WebSocket handshake...

FastGPT code issues and vulnerabilities

FastGPT is an open-source knowledge base question-answering system based on large language models developed by Labring. Versions of FastGPT prior to 4.14.9.5 contained code vulnerabilities. These vulnerabilities stemmed from the MCP tool endpoints in FastGPT, which accept URL parameters provided ...

CVE-2026-0560

A Server-Side Request Forgery SSRF vulnerability exists in parisneo/lollms versions prior to 2.2.0, specifically in the /api/files/export-content endpoint. The downloadimagetotemp function in backend/routers/files.py fails to validate user-controlled URLs, allowing attackers to make arbitrary HTT...

Incorrect Authorization

Overview openclaw is a 🦞 OpenClaw — Personal AI Assistant Affected versions of this package are vulnerable to Incorrect Authorization via the /v1/models HTTP endpoint, which does not enforce the required operator read scope. An attacker can access and enumerate model metadata by sending...

GHSA-68F8-9MHJ-H2MP OpenClaw has a Gateway HTTP /v1/models Route Bypasses Operator Read Scope

Fixed in OpenClaw 2026.3.24, the current shipping release. Summary The OpenAI-compatible HTTP endpoint /v1/models accepts bearer auth but does not enforce operator method scopes. In contrast, the WebSocket RPC path enforces operator.read for models.list. A caller connected with operator.approvals...

Server-side Request Forgery (SSRF)

Overview Affected versions of this package are vulnerable to Server-side Request Forgery SSRF via the /loadIG endpoint, which accepts user-supplied URLs and makes server-side HTTP requests without proper validation of hostnames, schemes, or domains. An attacker can probe internal network services...

CVE-2026-33992

pyLoad is a free and open-source download manager written in Python. Prior to version 0.5.0b3.dev97, PyLoad's download engine accepts arbitrary URLs without validation, enabling Server-Side Request Forgery SSRF attacks. An authenticated attacker can exploit this to access internal network service...

Spring AI 1.0.x < 1.0.4 / 1.1.x < 1.1.3 Multiple Vulnerabilities

The version of Spring AI installed on the remote host is 1.0.x prior to 1.0.4 or 1.1.x prior to 1.1.3. It is, therefore, affected by multiple vulnerabilities: - A JSONPath injection vulnerability in AbstractFilterExpressionConverter allows authenticated users to bypass metadata-based access...

PT-2026-31837

Fixed in OpenClaw 2026.3.24, the current shipping release. Summary The OpenAI-compatible HTTP endpoint /v1/models accepts bearer auth but does not enforce operator method scopes. In contrast, the WebSocket RPC path enforces operator.read for models.list. A caller connected with operator.approvals...